Tenofovir and Entecavir Are the Most Effective Antiviral Agents for Chronic Hepatitis B: A Systematic Review and Bayesian Meta-Analyses - pdf published attached
Download the PDF
Article in Press
Received 30 July 2009; accepted 4 June 2010. published online 21 June 2010.
Gloria Woo, George Tomlinson, # George Tomlinson, Yasunori Nishikawa, Matthew Kowgier, Morris Sherman, David K.H. Wong, Ba Pham, Wendy J. Ungar, Thomas R. Einarson, E. Jenny Heathcote, Murray Krahn
Toronto Health Economics and Technology Assessment Collaborative, University of Toronto, Toronto, Ontario, Canada
Departments of Medicine and Health Policy, Management and Evaluation, and Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada
University Health Network, Toronto General Research Institute and Clinical Studies Resource Centre Toronto Western Research Institute, Toronto, Ontario, Canada
Background & Aims:
The relative efficacies of licensed antiviral therapies for treatment-naive chronic hepatitis B (CHB) infection in randomized controlled trials have not been determined. We evaluated the relative efficacies of the first 12 months of CHB treatments.
Drugs evaluated were lamivudine, pegylated interferon, adefovir, entecavir, telbivudine, and tenofovir, as monotherapies and combination therapies, in treatment-naive individuals. Databases were searched for randomized controlled trials of the first 12 months of therapy in hepatitis B e antigen (HBeAg)-positive and/or HBeAg-negative patients with CHB published in English before October 31, 2009. Bayesian mixed treatment comparisons were used to calculate the odds ratios, including 95% credible intervals and predicted probabilities of surrogate outcomes to determine the relative effects of each treatment.
In HBeAg-positive patients, tenofovir was most effective in inducing undetectable levels of HBV DNA (predicted probability, 88%), normalization of alanine aminotransferase (ALT) levels (66%), HBeAg seroconversion (20%), and hepatitis B surface antigen loss (5%); it ranked third in histologic improvement of the liver (53%). Entecavir was most effective in improving liver histology (56%), second for inducing undetectable levels of HBV DNA (61%) and normalization of ALT levels (70%), and third in loss of hepatitis B surface antigen (1%). In HBeAg-negative patients, tenofovir was the most effective in inducing undetectable levels of HBV DNA (94%) and improving liver histology (65%); it ranked second for normalization of ALT levels (73%).
In the first year of treatment for CHB, tenofovir and entecavir are the most potent oral antiviral agents for HBeAg-positive patients; tenofovir is most effective for HBeAg-negative patients.
Abbreviations used in this paper: CHB, chronic hepatitis B, CrI, credible interval, RCTs, randomized controlled trials, MCT, mixed treatment comparisons, OR, odds ratio, PP, predicted probability.
An estimated 400 million people worldwide are chronically infected with the hepatitis B virus (HBV).1 Approximately 25% eventually will die of the liver-related complications of liver failure and hepatocellular carcinoma (HCC) if left untreated. Many chronically infected individuals, however, achieve spontaneous immune control of their HBV infection and do not require treatment. Loss of the viral protein marker, hepatitis B e antigen (HBeAg), frequently is associated with spontaneous immune control of HBV infection. Currently available antiviral therapies can suppress viral replication, whereas sustained immune suppression of HBV DNA is required to clear virus (loss of hepatitis B surface antigen [HBsAg]). For those who do not achieve spontaneous immune control, the goal is long-term suppression of HBV-DNA replication, which in some patients is followed by loss of HBsAg; the latter is associated with a lower risk of HCC and improved survival.2, 3
Currently available treatments include individualized single-agent therapy with interferon-alfa, nucleos(t)ide analogue polymerase inhibitors, and, potentially, combinations of these 2 forms of treatment. The specific drugs available worldwide for chronic hepatitis B (CHB) include standard and pegylated interferon alfa, lamivudine, adefovir, entecavir, telbivudine, and tenofovir. Interferon is used as a short-term treatment that, when successful, may lead to long-term immune control without the need for further antiviral therapy.4 Nucleos(t)ide analogues that directly inhibit the HBV reverse-transcriptase polymerase have no immune effect. Thus, once started, lifelong treatment may be required. Initiating therapy with each of these medications involves consideration of drug-specific trade-offs such as high and potentially lifelong medication costs, potential side effects including risks during pregnancy, and, perhaps most importantly, the risk of drug resistance over time.5 The risk of cross-drug resistance to polymerase inhibitors is a very serious problem because new HBV variants respond less well to new treatments with other polymerase inhibitors and often higher doses are required.6
Published studies evaluate the ability of drugs in treatment-naive CHB to achieve the following: (1) suppress HBV-DNA levels to clinically relevant levels (<1000 copies/mL, a level associated with inactive disease and a decreased risk of subsequent drug resistance)7, 8; (2) normalize ALT levels because normalization usually indicates cessation of ongoing liver injury; (3) induce HBeAg loss with seroconversion to anti-HBe because those who achieve this outcome may no longer require ongoing antiviral therapy; (4) decrease serum HBsAg titer because subsequent loss of HBsAg is a marker of sustained viral suppression; (5) improve liver histology; and (6) not to cause serious adverse events either while on or shortly after stopping therapy. Except for the trials of pegylated interferon, results are reported only for as long as the patients remain on treatment. Hence, few studies address the question of whether HBV-DNA suppression can be sustained long term once any of the oral therapies are discontinued.
The purpose of our study was to systematically review all published randomized controlled trials (RCTs) of drugs used to treat CHB as monotherapies or combination therapies to estimate their relative treatment efficacies at the end of 1 year of treatment and to rank the treatments according to the success rates for each outcome.
Many new antiviral treatments for CHB have become available within the past 2 decades. The first drug approved was interferon, an immune modulator and antiviral. After its inception, there has been a shift toward focusing on oral antiviral drugs, nucleos(t)ide analogues. RCTs comparing these treatments have been limited to comparing 2-3 drugs or drug combinations at a time whereas traditional meta-analytic techniques are limited to comparing 2 interventions. This has left clinicians to make their own judgments about the relative efficacy of treatments for which head-to-head trials are not available.
In our study, we used Bayesian MTC to evaluate the relative efficacy of all available treatments across 6 surrogate clinical outcomes. We consolidated the information of all RCTs that included the treatments of interest to provide the probability of an outcome at the end of 1 year of treatment as well as a rank for all treatments. The results of our analysis suggest that in treatment-naive individuals, entecavir and tenofovir are most effective at the end of the first year of therapy in HBeAg-positive patients whereas tenofovir is most effective in HBeAg-negative patients based on an overall assessment of all surrogates outcomes.
Our study focused on assessing surrogate clinical outcomes at the end of the first year of treatment. In recent years, there has been a shift toward evaluating outcomes that will reflect long-term improvements in the prognosis of those with CHB. Two authors (Fattovich et al2 and Hui et al3) have suggested that loss of HBsAg is the optimal goal of treatment and is the only surrogate marker of successful immunologic control and is associated with a lower incidence of cirrhosis and HCC and improved survival rates. Loss of HBsAg was not seen in those patients on oral antiviral treatment for 1 year but was seen in some patients who received pegylated interferon at 1 year after the initiation of treatment in the studies we reviewed. Although useful as an initial comparison of the various drugs, the utility of this review is limited to standard clinical practice in which the nucleos(t)ide analogue polymerase inhibitors are used and they are rarely prescribed for just 1 year. Rather, long-term, perhaps lifelong, suppressive therapy may be required. Important treatment issues such as long-term drug cost and drug toxicity, treatment failure owing to inadequate patient adherence, and/or drug resistance have not been addressed. In our study, few severe adverse events were reported in the first year of treatment. However, incidents of renal toxicity, lactic acidosis, neuropathy, as well as myositis have now been reported with long-term treatment using nucleos(t)ide analogues. Finally, the change in biomarker status examined in this review, such as loss of HBeAg or suppression of HBV DNA to undetectable levels, was determined while continuing antiviral therapy and the durability of this response on stopping therapy is not well known, particularly for the newer agents, tenofovir and entecavir.
Our study had limitations. First, the number of studies included for each pair-wise comparison was small. There was only one study evaluating the efficacy of tenofovir, one of the recommended treatments. The quality of reporting for this study was optimal; however, the power of the comparison was limited, hence the wide CrIs. In addition, there was a limited number of patients within each of the oral combination studies as well as a limited number of studies assessing oral combination therapy. Our analysis may suggest that oral combination therapy does not improve surrogate outcomes but studies of different combinations may provide different results. Our review also was limited to fully published studies in the English language. A number of clinical studies of CHB have been conducted in non-English speaking countries.
Other limitations include variation in definitions, measurements, patient characteristics, and protocols across studies and the quality of reported data. For example, the threshold values for undetectable HBV DNA varied significantly, thus earlier studies with higher thresholds tended to have higher proportions of subjects with undetectable HBV DNA. There were too few studies with any given threshold value to determine its effect on the OR for each pair-wise comparison; any variation in treatment effect owing to the threshold became part of the random between-study variance. We chose a threshold of less than 1000 copies/mL because studies have shown that for treatment with lamivudine, telbivudine, and adefovir, subsequent resistance is low for those whose viral load is maintained at less than 1000 copies/mL.36 In addition, the error of the diagnostic test is approximately 0.5 log copies so that a viral load of 300-1000 copies/mL is within the error of the test, and a majority of studies that used polymerase chain reaction to detect HBV-DNA levels were below that of 1000 copies/mL.
This study offers some insight into the relative benefits of current drugs at 1 year of treatment. Because many of these treatments will be taken for much longer, perhaps for a lifetime, these data are not sufficient to definitively resolve the question of optimal treatment choice. Although hard outcomes such as HCC, liver failure, and death are the most important clinical end points, these are rare events when observation times are only 1-2 years. There is presently no consensus regarding the most appropriate surrogate markers of a long-term outcome or even the validity of on-treatment measurements.
The controversy about which drug to use is dwarfed by the controversy about who should be treated. Patients with cirrhosis and ongoing viral replication appear to benefit from treatment in terms of rate of progression of liver disease. The only randomized trial conducted in patients with advanced liver disease was stopped prematurely because of a dramatic reduction in rates of liver failure and liver cancer.37 Any potential long-term benefits of treatment are unclear among patients with earlier stage disease.
Chronic infection with hepatitis B occurs predominantly in those who acquired the infection as a neonate or young infant. Neonatal infection induces an early immune-tolerant phase of the disease. Only when the infected individual enters the phase of active HBeAg-positive hepatitis is loss of HBeAg and seroconversion to anti-HBe possibly reducing the risk of developing significant chronic liver disease.3 This seroconversion when followed by sustained immune control has been shown, in recent publications of such individuals followed up for a 20- to 30-year period, to have a 50% chance of spontaneously losing HBsAg. Those who lose HBsAg before the age of 50 years have a reduced likelihood of HCC.38 On the other hand, individuals with HBeAg-positive hepatitis who fail to seroconvert within 3-6 months or who subsequently develop HBeAg-negative hepatitis may warrant lifelong antiviral therapy because of the risk of progressive liver disease and HCC.
The first impetus toward earlier treatment was provided by a study performed in men in Taiwan. This large, population-based study, which recruited men between the ages of 30-65 years, showed that HBeAg status was associated with a high 9-year risk of HCC.39 This study in men also noted that the risk of HCC was greatest in those who were older and who remained HBeAg-positive. The REVEAL study, which also was performed in Taiwan of both men and women between the ages of 30 and 65, showed that HCC risk was increased significantly in individuals with a high serum HBV-DNA level (>10,000 copies/mL).40 This study showed that high serum HBV-DNA level was a prominent risk factor independent of HBeAg status, serum ALT level, and the presence of cirrhosis. These studies suggest that induction of HBeAg seroconversion and effective control of viral replication after antiviral therapy would lower the risk of HCC.
Finally, we do not know when and if it is appropriate to stop oral antiviral therapy once started if the individual has not undergone HBeAg loss or seroconversion for HBeAg-positive hepatitis. We only know that individuals who spontaneously achieve sustained immune control and who lose HBsAg (particularly if this occurs when the patient is younger than age 50), do secure a better chance of survival. Because treatment may continue for many years, the potential benefits of antiviral therapy on liver-related morbidity and mortality must be carefully weighed against the possibility of future drug resistance, high lifetime costs, and adverse effects.
Our systematic review and Bayesian MTC analysis shows that for patients with HBeAg-positive CHB, entecavir and tenofovir are the most effective treatments, whereas for HBeAg-negative patients, tenofovir is the most effective treatment as measured by our defined surrogate clinical outcomes. (Supplementary Table 1, Supplementary Table 2).
Search Results and Study Characteristics
We initially identified 3338 potentially eligible citations. After evaluating these citations and their bibliographies, we included 20 trials9, 10, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 (Figure 2); 15 in HBeAg-positive patients, 8 in HBeAg-negative patients. Three of these studies evaluated both HBeAg-positive and HBeAg-negative patients.
Table 1 provides a summary of the characteristics of the 20 studies that met our inclusion criteria. Double-blinding was described fully in 12 studies, partially in 2 studies, 4 studies were open-label studies, and 2 studies did not report blinding. As assessed by the Cochrane Risk of Bias tool, inadequate sequence generation provided the largest risk of bias followed by inadequate allocation concealment (Figure 3).
For the treatment of CHB, there were 10 trials with 1540 individuals treated with lamivudine, the common comparator used for our analysis (Table 1). When outcomes are beneficial, an OR greater than 1 reflects a treatment that is more effective in comparison with the common comparator (lamivudine) whereas an OR less than 1 reflects a less effective treatment. In direct comparisons, placebo was significantly less effective in inducing ALT normalization (OR, 0.11; 95% CrI, 0.03-0.38) and improving liver histology (OR, 0.27; 95% CrI, 0.09-0.84) compared with lamivudine.
In indirect comparisons, lamivudine was superior to placebo for all surrogate outcomes except inducing HBsAg loss (Supplementary Table 1).
Pegylated interferon (n = 407) was evaluated as monotherapy in 2 trials. Studies of standard interferon were omitted because this short-acting form of interferon is no longer considered by most as the standard of care. Direct comparisons suggested that it is significantly more effective than lamivudine monotherapy in inducing decreases in HBeAg loss and HBsAg loss. Pegylated interferon ranked among the top 4 treatments for HBeAg seroconversion (predicted probability [PP], 0.23; 95% CrI, 0.14-0.35), HBeAg loss (PP, 0.33; 95% CrI, 0.15-0.54), HBsAg loss (PP, 0.01; 95% CrI, 0-0.07), and histologic improvement of the liver (PP, 0.52; 95% CrI, 0.06-0.95) (Table 2).
Adefovir (n = 337) was evaluated in 4 trials and was not significantly better than lamivudine. Adefovir did not rank above fourth place for any outcome.
Entecavir (n = 408) was a comparator in 3 trials. In direct comparisons, it had increased efficacy in comparison with lamivudine in improving liver histology (OR, 1.56; 95% CrI, 1.12-2.19). Entecavir consistently ranked in the top 5 treatments for all surrogate outcomes and was ranked first with regard to improving liver histology (PP, 0.56; 95% CrI, 0.12-0.94).
Telbivudine (n = 684) was a comparator in 4 trials. In direct comparisons, it had improved efficacy compared with lamivudine in inducing undetectable HBV DNA (OR, 2.34; 95% CrI, 1.31-5.36) and histologic improvement of the liver (OR, 1.41; 95% CrI, 1.09-1.84). Indirect comparisons confirmed the results of the direct comparison of HBV-DNA undetectability. Telbivudine's rankings ranged from second for HBeAg loss to last for HBsAg loss.
Tenofovir (n = 176) was a comparator in one study. In indirect comparisons, tenofovir showed improved efficacy compared with lamivudine in inducing undetectable HBV-DNA levels (OR, 23.34; 95% CrI, 6.19-76.39). Tenofovir was consistently ranked in the top 3 treatments for all surrogate outcomes except HBeAg loss, for which no data were available. It was ranked first for inducing undetectable HBV DNA (PP, 0.88; 95% CrI, 0.69-0.97) normalization of ALT levels (PP, 0.66; 95% CrI, 0.41-0.91), HBeAg seroconversion (PP, 0.20; 95% CrI, 0.07-0.43), and HBsAg loss (PP, 0.05; 95% CrI, 0-0.54).
Three combination strategies were assessed in this analysis, lamivudine plus pegylated interferon (n = 451), lamivudine plus telbivudine (n = 41), and lamivudine plus adefovir (n = 53). In indirect comparisons, lamivudine plus pegylated interferon was more effective in inducing undetectable HBV DNA than lamivudine alone (OR, 3.08; 95% CrI, 1.88-4.91). In overall rankings, this combination was first in inducing HBeAg loss (PP, 0.39; 95% CrI, 0.18-0.63) and third for HBeAg seroconversion and second for HBsAg loss. In neither direct nor indirect comparisons were significant improvements found with combination therapy of 2 oral therapies (ie, lamivudine plus telbivudine or lamivudine plus adefovir) relative to lamivudine monotherapy.
The all-oral antiviral combinations of lamivudine plus telbivudine and lamivudine plus adefovir were ranked low in comparison with other therapies.
The between-study standard deviations of log-ORs for the surrogate outcomes, undetectable HBV DNA, ALT normalization, HBeAg seroconversion, HBeAg loss, HBsAg loss, and histologic improvement had posterior medians of 0.14, 0.29, 0.16, 0.27, 0.58, and 0.30, respectively (Table 3).
In indirect comparisons, lamivudine (n = 740) was more effective in comparison with placebo in inducing undetectable HBV DNA. In comparison with other treatments, lamivudine was ranked in the bottom 2 treatments for all outcomes measured (Table 4).
In direct comparisons, pegylated interferon was less effective than lamivudine in inducing undetectable HBV-DNA levels and ALT normalization 1 year after the initiation of therapy. In direct pair-wise comparisons with lamivudine, neither adefovir, telbivudine, entecavir nor tenofovir were more efficacious. However, in indirect comparisons, treatment with entecavir was more efficacious for all outcomes. Entecavir was ranked among the top 4 treatments for all outcomes, HBV DNA (PP, 0.88; 95% CrI, 0.65-0.97), ALT normalization (PP, 0.76; 95% CrI, 0.25-0.98), and histologic improvement (PP, 0.64; 95% CrI, 0.01-1.00). Tenofovir ranked first for HBV-DNA suppression (PP, 0.94; 95% CrI, 0.56-1.00) and histologic improvement (PP, 0.65; 95% CrI, 0.01-1.00), and second for ALT normalization (PP, 0.73; 95% CrI, 0.07-1.00) (Supplementary Table 2).
Lamivudine plus pegylated interferon (n = 296) was more effective than lamivudine alone in inducing undetectable HBV-DNA levels (OR, 2.40; 95% CrI, 1.41-4.19). However, it was less effective in inducing ALT normalization (OR, 0.35; 95% CrI, 0.23-0.55) at 1 year.
The standard deviation for the surrogate outcomes, undetectable HBV DNA, ALT normalization, and histologic improvement, were 0.48, 0.83, and 0.48, respectively (Table 3).
Severe Adverse Events
Severe adverse events were documented inconsistently and had varied definitions for each study, which prevented quantitative analysis. The greatest number of events occurred with monotherapy and combination therapies involving pegylated interferon.21, 22 Most events resolved after a decrease in dosage of pegylated interferon, withholding of dosages for a short period of time, or termination of therapy.21
Depression was reported as the main concern in patients treated with pegylated interferon.24, 32 The rate of depression on pegylated interferon therapy was 5%; combination therapy of lamivudine plus pegylated interferon was associated with similar depression rates (6%-7%).24, 32 The reported average rates of discontinuation of therapy were as follows: pegylated interferon, less than 5%; lamivudine, 2.8%; adefovir, 1.0%; entecavir, 1.8%; and tenofovir 1.0%. The most common adverse events reported while on treatment with oral antivirals were headache, upper respiratory infection, nasopharyngitis, cough, fatigue, upper abdominal pain, back pain, and diarrhea, most of which were mild-to-moderate in severity as reported by each of the studies.20, 27, 30, 34, 35 There was inconsistent documentation for adverse events after discontinuation of therapy. All treatments induced low rates of grade 3 or 4 changes in clinical laboratory values of liver tests (serum ALT, creatine kinase), and the rates were similar for each treatment.28, 34, 35 The rates of hepatic flares on therapy were as follows: lamivudine (4%30), pegylated interferon (8%32), adefovir (2%32), entecavir (1%10), telbivudine (1%34), tenofovir (6%34), and combination therapy with lamivudine plus adefovir (7%30).